Рад у зборнику

објављена

енглески

Danka Jokić, Ranka Stanković, Cvetana Krstev, Branislava Šandrih

A Twitter Corpus and Lexicon for Abusive Speech Detection in Serbian

3rd Conference on Language, Data and Knowledge (LDK 2021)

Gromann, D. and Serasset, G. et al.

MDPI AG

2021

Uvredljivi govor na društvenim medijima, uključujući psovke, pogrdni govor i govor mržnje, dostigao je nivo pandemije. Sistem koji bi bio u stanju da detektuje takve tekstove mogao bi da pomogne da internet i društveni mediji postanu bolji virtuelni prostor sa više poštovanja. Istraživanja i komercijalna primena u ovoj oblasti do sada su bili fokusirani uglavnom na engleski jezik. Ovaj rad predstavlja rad na izgradnji AbCoSER-a, prvog korpusa uvredljivog govora na srpskom jeziku. Korpus se sastoji od 6.436 ručno označenih tvitova, od kojih je 1.416 označeno kao tvitovi koristeći neku vrstu uvredljivog govora. Tih 1.416 tvitova je dalje podklasifikovano, na primer na one koji koriste vulgaran, govor mržnje, pogrdni jezik, itd. U ovom radu objašnjavamo proces prikupljanja podataka, beleženja i izgradnje korpusa. Takođe razmatramo rezultate početne analize kvaliteta napomene. Na kraju, predstavljamo strukturu leksikona uvredljivog govora i njeno obogaćivanje uvredljivim okidačima izvučenim iz AbCoSER skupa podataka.

Abusive speech in social media, including profanities, derogatory and hate speech, has reached the level of a pandemic. A system that would be able to detect such texts could help in making the Internet and social media a better and more respectful virtual space. Research and commercial application in this area were so far focused mainly on the English language. This paper presents the work on building AbCoSER, the first corpus of abusive speech in Serbian. The corpus consists of 6,436 manually annotated tweets, out of which 1,416 were labelled as tweets using some kind of abusive speech. Those 1,416 tweets were further sub-classified, for instance to those using vulgar, hate speech, derogatory language, etc. In this paper, we explain the process of data acquisition, annotation, and corpus construction. We also discuss the results of an initial analysis of the annotation quality. Finally, we present an abusive speech lexicon structure and its enrichment with abusive triggers extracted from the AbCoSER dataset.

13:1

13:17

17

10.4230/OASIcs.LDK.2021.13

978-3-95977-199-3

2190-6807

uvredljivi jezik, govor mržnje, srpski, tviter, leksikon, korpus

abusive language, hate speech, Serbian, Twitter, lexicon, corpus

https://www.mdpi.com/2076-3417/11/7/2892/pdf

М30

М33

Отворени приступ

Creative Commons – Attribution-Share Alike 4.0 International

.pdf